The harness landed in v1.30.35, after the slices that built this
release were already cut — so the reaction line's prompt module was
never registered, and its own docblock's warning against the
two-locale collapse was untested. Registering it surfaced a real gap:
the meaning-first instruction lived only in the system prompt's own
words, never in the user prompt in the family's shared phrasing, which
is where the rest of the surfaces carry it and where the tone contract
checks for it. The German body carried no such instruction at all.
Threads locale through the user-prompt builder so the fixed
instruction line matches the reader's own reply language, and wires
the per-user per-day opener-archetype rotation the milestone spec
called for and the first implementation had not connected.
Nothing warms this queue and no read path triggers it, so an unregistered name
fails open in the worst way: every send resolves, every job vanishes, and the
symptom is indistinguishable from the surface's honest empty state.
The paragraph is model output. Rendering it as markup would be an XSS surface,
and this project ships no markdown library so the question stays closed — the
test is what keeps that from being treated as a formatting decision.
The one-paragraph-per-workout claim is a unique index, the cascade is an FK,
and the daily-cap window is a SQL fact — a mocked upsert reports success for
all three regardless.
The projection's central claim was carried by a comment. It is now walked
before it is returned: any non-numeric leaf at any depth throws, and the only
two string leaves are allowlisted by shape — the sport type against the schema
enum itself, and the day key against its format.
This matters because sport_type is a TEXT column whose API write path is
enum-constrained but whose backup-restore path accepts any non-empty string, so
a restored row can carry arbitrary text. It is re-asserted at the projection
boundary rather than trusted from the column. The heart-rate provenance string
is dropped rather than allowlisted, and route geometry stays consumed-not-carried
— it yields one number, and location history may not ride a prompt.
Own-history rows now collapse cross-source twins through the canonical picker,
as the detail page's comparison already did; a session recorded by two paired
watches was inflating the sample size and biasing every median.
Four suites, each mutation-checked against a reverted protection:
- route-workout-scope: the workout row is read exactly once across a ten-turn
conversation that keeps sending the id, the where clause always names
userId, another user's id resolves to nothing and the chat proceeds, and
the section sits inside the health-data fence.
- workout-evidence: the numbers-only closure. An instruction-shaped
sport_type — reachable through the restore path, which types the column as
free text — folds to "other" and never reaches the block; a free-text leaf
at any depth throws.
- page-launch-params: an unknown ?scope= is ignored silently, and ?workout=
yields to both ?c= and the fenced ?doc= path.
- composite-score-anatomy-coach: every score sheet carries an opener, a
scope narrowed to the composite's inputs, and auto-send.
resolveLaunchState's shape assertion gains the workoutId field.
Three properties, each with a mutation-verified guard:
The chip and the line run on DIFFERENT clocks and must keep doing so. The chip
expires after three hours; the line stands for the rest of the local day.
Collapsing them onto one clock reads like a simplification and silently deletes
the day's read at lunchtime.
Degradation is the headline. For every reason a line can fail to exist — no
provider, no consent, an exhausted budget, a dead provider, an unusable output
— the row stays line-less, the reservation is reconciled, and the chip and the
provisional-to-final flip still work. The degraded digest is built through the
real composer rather than a hand-shaped fixture, so the assertions cannot agree
with themselves.
The server pass carries no wall-clock time. The hero paints on the SSR pass
because it is the LCP element, so a locale-formatted time inside it would differ
from the browser's — React #418. The tests render through the server snapshot,
which is exactly the pass that must stay clock-free.
The seeded end-to-end spec lands a sleep batch through the real ingest route,
writes the arrival marker, and asserts the chip and the phase flip in place on
a refetch trigger the digest hook already declares — no reload, no new cadence.
One test per gate, each asserting the refusal reason AND that the provider
entry was never reached — a gate that refuses but still resolves a provider
would pass a reason-only assertion while costing what the gate exists to
prevent. The cap and the hash are asserted independently of each other, since
the singleton queue key that normally collapses a burst is best-effort.
The backfill fixture walks the real chain: 750 historical sessions plus 40
re-synced ones plus one fresh, through the live classifier and the live spine
worker, yielding exactly one generation.
The peak detector changed under test: on a flat ride the ninetieth percentile
sits a beat above the mean, so a bpm of sampling noise read as two efforts the
rider did not do. Peaks now need an absolute prominence over the session mean
as well.
The spine dispatches a generator-free enqueue; the worker resolves the
provider. Five gates stand in front of it in ascending cost order — both
modules on, ten-minute duration floor, a hard count of four a day in the
user's own timezone, an input hash over the deterministic evidence, and the
token ledger inside the shared status chokepoint. Each refusal returns a
status rather than throwing, because retrying against a ceiling that does not
move until midnight is an unbounded loop.
The detail route only reads: a workout with no row serves null and renders no
card, which is every workout that predates this and every re-synced one. An
undecryptable paragraph degrades to no card rather than failing the page.
Today gains two additive-nullable digest fields. `justIn` names the day's
newest arrival while it is still news — under three hours old, after which the
record is simply current rather than new. `reactionLine` carries a single
generated sentence about that arrival and replaces the hero's standing lead for
the rest of the local day, never appending a second paragraph, so the hero stays
one lead line tall and nothing below it shifts.
Nothing here polls. The open dashboard picks the change up on the cadence it
already runs — a state change, not a notification.
The generated sentence is bounded before it is written: the arrival spine's
unique row per kind per local day IS the throttle, the worker re-checks the
same claim before spending, and the reservation is reconciled on every exit
including each failure path. Without a provider, without consent, or against an
exhausted budget the row stays line-less and the surface falls back to the
deterministic lead — the chip and the provisional-to-final flip are unaffected.
The chip's time string is formatted client-side after mount; `justIn.at`
crosses the wire as an ISO instant so the server never renders a local
wall-clock time the browser would disagree with.
Status cards and derived-score sheets read one way: they state a verdict and
end there. The metric pages already carried the hand-off; the score sheets
passed neither an opener nor a scope, so they painted no action at all.
- Score sheets (sleep / readiness / recovery / stress / strain) gain the same
auto-sent opener the metric pages use, each scoped to the inputs that drive
the composite rather than to a block it has no equivalent of.
- Workout detail's coach launch carries the workout itself. The chat request
gains an optional workoutId; on the FIRST turn the route narrows by
{ id, userId } and pins ONE bounded evidence section onto the snapshot.
Later turns reuse the pinned snapshot, so per-turn work stays constant in
conversation length.
- /coach?workout=<id> covers the full-page and maximize paths.
The workout block deliberately does not take the fenced transport the stored-
document path takes: a workout row is the user's own sensor record, and the
projection admits no free text. That is enforced, not assumed — sport_type is
a TEXT column, so it is re-asserted against the closed sport union, metadata
is excluded apart from its numeric zone durations read through a Zod slice,
and a recursive guard throws on any non-numeric leaf.
Own-history sport averages move to src/lib/workouts/sport-context.ts so the
narrative and the detail page cannot quote different figures.
The evidence block is a numbers-only server projection: zone seconds, the
workout-window heart-rate shape, route climb and a same-sport 90-day own-history
median. Free-text metadata never enters it, and the sport string is narrowed
against the closed enum first, so the payload carries no text the user did not
generate through a measured field.
The prompt composes the shared assessment base and adds the session contract:
device-attributed, own-history-comparative, effort acknowledged where the
numbers show it, and no training prescriptions of any kind. Registered in the
tone harness beside the assessment cards.
Its own table, cascading from the workout row, because workout deletes are
hard deletes. The paragraph is AES-256-GCM at rest and registered for key
rotation; the input hash and prompt version make a re-post of the same
session a free no-op.
The deterministic fallbacks nine metric surfaces render when no provider is
reachable carried German and English bodies only. `getLocalizedText` returned
`locale === "de" ? de : en`, so French, Spanish, Italian and Polish readers got
English — on the no-key path, an outage, a spent budget or a timeout, which is
precisely when this text is the whole assessment.
Every string now has a body per locale: the verdict leads and value sentences
of the signal-grounded composer, the per-metric labels and pointers, and the
eight no-signal floors. Selection reads the reader's locale; English stays as
the structural floor for a locale that is ever genuinely missing a body.
The bodies are written per language, not translated word-for-word. fr/es/it
replace the comparative adjective ("higher than your usual average") with a
preposition — au-dessus de / por encima de / sopra — because a comparative
would have to agree in gender with the metric noun, and the labels do not share
one. Polish restructures further: the sentence takes `wynosi`, which is
invariant for the subject's gender, and the comparison uses powyżej / poniżej
with the genitive; the labels span all three genders, so nothing that inflects
can sit in a shared template.
The tone harness grades the four locales by the same rules rather than passing
them by default: false cheer, clinical-verdict and imperative-opener vocabulary
extended to all six, the no-assessment sentinel keyed per locale instead of a
de/en binary, and a byte-identity guard that fails if any locale silently falls
back to the English text again. The value-led-opener rule was rewritten as
well — it assumed the copula sits against the figure, which under-matched its
own second locale (German "liegt aktuell bei 24") and would have missed all
four new ones.
The batch route now also resolves the user's timezone and module map through
the arrival seam, so a raw call count on user.findUnique no longer isolates
the read this guard is about. Filtering by the SELECT shape keeps the
once-per-batch contract pinned while letting the unrelated cached lookups
through.
The no-signal floors were the coldest text the product renders. They run when
there is nothing to ground against, and they opened on a bare order — "Measure
blood pressure at rest and under comparable conditions", "Evaluate mood trends
over several weeks" — so the one moment the surface has nothing to say was also
the one moment it told the reader what to do.
They now open on what the surface can honestly say: no assessment is available
right now. That is the meaning-first opening for this case, and it is the only
one available, because a floor with no signal may not imply a read. What
follows describes how the measure behaves rather than ordering the reader
around. The signal-grounded lines above them are unchanged.
The prompt version bump re-fingerprints the assessments cached under the old
biomarker prompt; the per-card input hash folds the constant in, so the
regeneration is one-time and paced by the nightly warm.
Backfill immunity is proved by replay, not by a hand-picked case: a ten-year
Apple Health export (12772 writes) and a 30-day provider re-sync (58 writes)
must both emit zero events, while a fresh sleep night, workout and first
weight of the day emit exactly one each.
The provider-isolation guard walks the worker's static import graph and fails
if any provider client or resolver becomes reachable. It carries two positive
controls, so a broken entry path cannot make it pass against an empty graph.
The integration test covers what unit mocks cannot: the unique index, the
ON CONFLICT DO NOTHING claim, the losing side of ten concurrent claims, the
forward-only marker, ciphertext at rest, the delete cascade, and the
retention window.
Every ingest path in the product can now emit a typed, salience-classified
event into pg-boss, consumed by one worker that claims a durable per-day
reaction marker and fans out to the reaction surfaces.
The recency classifier runs at the seam, before anything is enqueued, so a
mass backfill — a multi-year Apple Health export, a provider re-sync, a
30-day catch-up — emits nothing at all. Non-salient writes are annotated and
dropped.
The worker makes no provider call by construction: it imports no provider
module, writes only markers, and enqueues nothing that is not budget-gated
downstream. Business refusals return a status rather than throwing, so a
refusal is never retried against a ceiling that will not move until the local
day rolls over.
Seams: sleep (one hook, inherited by all seven transports), workouts (the
batch route, the four provider syncs, the export importer), weight and blood
pressure (both measurement routes), labs (manual, OCR commit, document fact
commit).
The provider workout syncs could not previously tell a created row from an
updated one, so a re-sync of a recent session would have re-fired on every
poll; they now recover the distinction from the row the upsert already
returns.
Reaction lines are AES-256-GCM at rest, following the InsightNarrative
precedent, and the column is registered for key rotation. Markers are pruned
after 14 days by a daily maintenance job.
A deterministic grader over the assembled assessment prompts and the
deterministic fallback text. No provider, no network, no database — it runs in
the normal suite in under half a second, which is the point: the live-judge
eval needs a key and can only advise, so the regressions that actually shipped
need a check that can block.
Three classes are blocking. A surface stops leading with meaning (it drops the
shared opening contract, or its user prompt loses the meaning-first
instruction). A locale collapses (fr/es/it/pl fall back to the German body or
lose the reply-language directive). A prompt reintroduces a value-first
instruction — matched with negation handling, since the shared prompts quote
that failure mode in order to ban it.
The deterministic fallbacks are graded on the same axis, including the
asymmetry that matters: a signal with no baseline supports no confident read,
so that one line honestly opens on the value. Pinning it stops a future
"always lead with meaning" edit from pushing the surface into claiming
something the data does not support.
A coverage guard fails when an assessment prompt module has no harness entry,
so the next surface cannot stay uncovered the way the biomarker card did.
The per-biomarker assessment was the last surface in the family building its
prompt inline: a five-line scaffold that skipped the base system prompt, took
no opener hint, and instructed the model to "state the current value" — so the
one card where a bare number reads most like a verdict was value-first by
instruction while every sibling had moved to meaning-first.
It now composes exactly as the seven bespoke cards do. The new prompt module
carries a METRIC section that states the one thing this card needs and the
others do not: its snapshot has no signal block, so the read comes from the
pre-computed range placement plus the direction across draws.
Also point both archetype-driven cards at a registered budget. The generic
metric card and the biomarker card each carried temperature/maxTokens inline
at 0.45/1000, so the registry that documents every surface's ceiling was
silently wrong about the two widest assessment surfaces.
The single-marker no-key fallback replaces the multi-metric overview tip the
card used to render — text about watching several metrics at once, on a card
showing one lab value.
Signing every other device out, revoking a single session, revoking a
trusted device and revoking them all each fired on a single tap. None is
undoable — the other devices simply have to sign in again, and a revoked
trusted device asks for its second factor next time — while every other
destructive surface in the app confirms first. On a shared screen a
mis-tap is a support call.
A sweep of the rest of the tree found two more hard deletes with neither
a confirmation nor an undo: a Coach reminder and a travel entry. Both now
confirm. Deleting a document deliberately does not: it soft-deletes and
offers an undo, which covers the mis-tap.
The four controls share one ConfirmButton rather than four copies of the
dialog, so the next destructive control gets the confirmation by using
it instead of by someone remembering. Copy is per control and states the
actual consequence — no claim of more finality than is true.
The widget catalogue grew to 42 ids, so the snapshot round-trip count
follows. The mood card's unit key carries "/5" rather than an empty
string — the locale-integrity guard forbids empty values, and the scale
is the honest thing to print next to a 1-to-5 score. It joins the
identical-across-locales allowlist alongside the other unit notations.
The native pen list renders these two as its headline and subhead. The
server held neither, so a container registered on the web could not be
shown there without fabricating an empty card, and the client kept its
own copy locally with no back-fill possible.
Both columns are additive and nullable (migration 0255). Every existing
row is truthfully NULL — the values were never collected, and a plain
supply row such as a blister pack or a bottle has no use for either.
Strength stays free text rather than a parsed number + unit: pens state
it per dose, per ml, or per cartridge depending on the product, and
forcing a numeric split would lose which of the three was printed.
On PATCH an omitted field leaves the stored value untouched, the same
rule notes already followed, so a client that does not know the fields
cannot blank what another populated; an explicit null still clears. The
web add-container form collects both for a PEN, so the columns do not
ship empty forever.
Three surfaces where the words no longer matched the code, plus the missing
web control for a preference the server already honours.
Delete confirmations. Four dialogs promised permanence on surfaces that
render an Undo: the measurements and mood bulk bodies, the lab-result body,
and the illness-episode body (which said "cannot be undone" sixty lines
below the code that fires the Undo toast). All four now use the same
sentence the single-row measurement and mood dialogs already used. A sweep
of every other confirmation in the tree found the rest accurate — the
biomarker cascade, the mood-tag purge, and the account purge keep their
warning because those deletes really are unrecoverable.
Sleep debt. v1.19.0 replaced the summed shortfall with a running balance
that credits catch-up sleep; the labels still described the old sum, so the
word contradicted the number for anyone who slept in after a short night.
Corrected in all six locales, plus the places the old semantics had
propagated: the Coach system prompt, the OpenAPI field descriptions the
native client reads, and four module docblocks.
Glucose reference band. `hasDiabetes` selects the tighter reference band and
grounds the Coach, but no web surface set it — a web-only account was pinned
to the non-diabetic scale with no way out. The control goes on Targets,
above the per-metric editor, because it answers the same question that page
answers and the two are ordered: it picks the default band, an override
below beats it. The copy says what it is — a reference band, not a
diagnosis, never inferred from a reading, no clinical action taken on it.
The passkey list was reported as always reading "never used". It does not
reproduce: the column is selected by the list route and stamped on every
verified assertion, both since v1.23.0. Pinned with tests on both halves so
the contract cannot quietly regress.
Both metrics exist in the record but the summary route emitted neither,
so a native client could render every home-screen tile except those two.
Mood reads the canonical daily series (`buildMoodDailySeries`) — the same
rollup-backed aggregation the web mood analytics and the dashboard
snapshot render — and emits the newest day's score with a trailing
seven-day sparkline. BMI comes from `computeBmi`, the one implementation
the web vitals dashboard already renders through the derived-metric
dispatch; the route resolves the figure so no client re-derives weight
over height squared.
Mood gates on the mood module through a synthetic `MOOD_ENTRY` key, the
same route `NUTRIENT_WATER` takes, because mood is not a measurement
type; without the join the card would ship past a disabled module. BMI
belongs to no module and rides through with the other core vitals.
The BMI card is omitted rather than emitted empty when the profile has no
height or no weight in the window: the card shape carries no honest
"unknown" state, and a null value would read as a logging gap rather than
a missing profile field.
The layout PUT preserved chartOverlayPrefs, selectedScoreRings and
heroRingOrder when a client omitted them, but the top-level
comparisonBaseline was not in that merge. serializeDashboardLayout clamps
a missing baseline to "none", so a layout save from a client that does
not know the field — the native client treats it as web-only and never
sends it — silently reset the comparison the user had picked on the web.
No error, no audit row.
Replace the hand-maintained three-field merge with an explicit
disposition map over every top-level layout field. The map is declared
`satisfies Record<keyof DashboardLayout, ...>`, so adding a layout field
without deciding replace-or-preserve fails typecheck; the preserve set
and the merge are derived from it rather than restated. The stored-layout
read is skipped when the client sent the full set.
Also catalogue the four clinical signals the native client pins — grip
strength, pain NRS, waist circumference, waist-to-height. They reached
the route as unknown ids, were filtered out ahead of Zod, and vanished
from the persisted layout, so a tile placement was lost on every save.
They join the iOS-only set because the web dashboard has no render path
for them; their charts live on the insights sub-pages. Raise the widget
array cap to 50 to clear the now-42-id catalogue.
Marc-André Bombeck